# Discriminant analysis for the validation of cluster analysis

I did a discriminant analysis for the validation of my cluster analysis. The cluster analysis is based on a PCA, so I used the components as the independant variables in the discriminant analysis. My question is: is there a restriction considering the number of the independant variables (components) compared to the number of my (cluster) groups?

Thanks a lot!

• How can DA validate the clusters? Please describe how you reason it in more details. – ttnphns Oct 30 '13 at 17:32
• Because it is frequently used in literature. – cathy Oct 31 '13 at 3:57

$Cluster=Component_1+Component_2,...,Component_n$